| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Fix python interpreter when installing in venv https://review.opendev.org/c/openstack/bifrost/+/949518 | 06:53 |
|---|---|---|
| rpittau | good morning ironic! o/ | 06:53 |
| queensly[m] | Good morning | 08:17 |
| abongale | Good Morning Ironic! | 08:39 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Auto-detect the RPC executor instead of hardcoding eventlet https://review.opendev.org/c/openstack/ironic/+/949593 | 10:56 |
| opendevreview | Dmitry Tantsur proposed openstack/ironic master: Auto-detect the RPC executor instead of hardcoding eventlet https://review.opendev.org/c/openstack/ironic/+/949593 | 10:57 |
| opendevreview | cid proposed openstack/ironic master: Fix redfish driver URL parsing https://review.opendev.org/c/openstack/ironic/+/949597 | 11:09 |
| opendevreview | Elod Illes proposed openstack/ironic unmaintained/2023.1: OSSA-2025-001: Disallow unsafe image file:// paths https://review.opendev.org/c/openstack/ironic/+/949182 | 11:10 |
| opendevreview | cid proposed openstack/ironic master: Fix redfish driver URL parsing https://review.opendev.org/c/openstack/ironic/+/949597 | 11:11 |
| opendevreview | Verification of a change to openstack/ironic master failed: De-duplicate two WSGIService's https://review.opendev.org/c/openstack/ironic/+/945480 | 11:11 |
| opendevreview | cid proposed openstack/ironic master: Fix redfish driver URL parsing https://review.opendev.org/c/openstack/ironic/+/949597 | 11:15 |
| opendevreview | Elod Illes proposed openstack/ironic unmaintained/yoga: OSSA-2025-001: Disallow unsafe image file:// paths https://review.opendev.org/c/openstack/ironic/+/949177 | 11:23 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: [WIP] use Pyhon 3.12 on centos9 https://review.opendev.org/c/openstack/bifrost/+/949449 | 11:57 |
| *** tosky_ is now known as tosky | 12:27 | |
| TheJulia | good morning | 12:57 |
| dtantsur | morning TheJulia | 13:00 |
| dtantsur | cid, JayF, TheJulia: a lot of musings about eventlet: https://etherpad.opendev.org/p/ironic-eventlet-removal | 13:00 |
| opendevreview | Habeeb Babasulaiman proposed openstack/bifrost master: doc: remove note that is not needed https://review.opendev.org/c/openstack/bifrost/+/947600 | 13:14 |
| opendevreview | Habeeb Babasulaiman proposed openstack/bifrost master: bug: drop --develop flag from testenv usage https://review.opendev.org/c/openstack/bifrost/+/946724 | 13:29 |
| opendevreview | Habeeb Babasulaiman proposed openstack/bifrost master: bug: drop baremetal introspection mention https://review.opendev.org/c/openstack/bifrost/+/946726 | 13:29 |
| opendevreview | Habeeb Babasulaiman proposed openstack/bifrost master: doc: remove note that is not needed https://review.opendev.org/c/openstack/bifrost/+/947600 | 13:29 |
| TheJulia | lower hanging fruit, but also impacting the signal to noise is some of the reliance in tests over the fun which using mock with time.sleep where eventlet sort of help solves that issue. After calling it a day yesterday, I ended up starting to toy with a removal, looks pretty good to just clean it up until we hit the redfish driver power code. :\ | 13:36 |
| dtantsur | oh yeah, the power code.. | 13:36 |
| * TheJulia glares at redfish/power.py | 13:37 | |
| TheJulia | ........... why do you still take 50 seconds | 13:37 |
| dtantsur | TheJulia: whatever we use for retrying is very hard, nearly impossible to mock | 13:38 |
| TheJulia | I know, there are some stupidly not fun test issues where eventlet sort of made it easy | 13:39 |
| TheJulia | ... There are some things we're just going to have to roll with some faith and other things we may need to refactor/decompose into smaller chunks in the end | 13:39 |
| dtantsur | true | 13:40 |
| dtantsur | feel free to update the document with any specific findings | 13:40 |
| TheJulia | I'm just working on a patch to try and rip it from tests, i figured that is likely the best path since we'll also be best served to let such code run for a while during the cycle | 13:41 |
| TheJulia | Anyhow, nearly time to walk the corgi overlord before time to gear up for a fun two hour meeting | 13:41 |
| dtantsur | fun or "fun"? :) | 13:41 |
| * dtantsur sends his regards to the corgi overlord | 13:42 | |
| TheJulia | heh, I just heard him climb into the human bed | 13:43 |
| TheJulia | I think he wants to sleep in a little longer | 13:43 |
| dtantsur | don't we all? | 13:43 |
| TheJulia | yes | 13:43 |
| TheJulia | Added a note under other uses | 13:50 |
| dtantsur | ++ | 13:55 |
| dtantsur | TheJulia: thanks! I'm also curious about your feedback on my thread pool sizing calculations | 13:56 |
| TheJulia | I've not entirely digested it, but I know some operators who are already overriding our existing default to be... as you set as an example, 300 | 13:58 |
| TheJulia | ugh, the issue with this test is is exercising the conductor utils code in an unmocked fashion so eventlet waitall was hiding it | 14:03 |
| TheJulia | and test now executes in 0.097 seconds | 14:03 |
| TheJulia | as opposed to 50-something | 14:03 |
| dtantsur | \o/ | 14:03 |
| opendevreview | Merged openstack/ironic master: De-duplicate two WSGIService's https://review.opendev.org/c/openstack/ironic/+/945480 | 14:04 |
| TheJulia | I suspect this is a pattern we've gotten into which is likely not great | 14:04 |
| TheJulia | The other fun issue is removing the monkey patch will likely require more test work regardless of what we do. But that is okay | 14:09 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: Fix python interpreter when installing in venv https://review.opendev.org/c/openstack/bifrost/+/949518 | 14:09 |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: [WIP] use Pyhon 3.12 on centos9 https://review.opendev.org/c/openstack/bifrost/+/949449 | 14:10 |
| drannou | Hello o/. Does anyone recently play with ironic-prometheus-exporter ? We are trying to use it, but we are not able to make it connect to the rabbit os notification ? From our investigations the conductor is correctly sending data in rabbit, but the exporter never connect to the queue | 15:22 |
| drannou | We are trying to start it in 'dev mode' like documentation is telling (export ... ; python -m flask run) | 15:23 |
| TheJulia | Not recently | 15:32 |
| dtantsur | drannou: the exporter cannot work with rabbitmq, it needs to use a fake local bus | 15:48 |
| dtantsur | cc iurygregory for more details | 15:48 |
| iurygregory | yeah, it wont connect to rabbitmq | 15:49 |
| drannou | ok that may explain why.... :) Might be a good idea to enforce that in the documentation ? | 15:50 |
| drannou | Do you have more explanation ? I was thinking to create an other python application that would just do the rabbit connexion and write the files, in order to keep Flask to desserve prom | 15:51 |
| iurygregory | we do have in the docs let me find | 15:53 |
| iurygregory | https://docs.openstack.org/ironic-prometheus-exporter/latest/configuration.html | 15:54 |
| iurygregory | it should be fake:// | 15:54 |
| dtantsur | drannou: I guess because nobody bothered with a proper bus support.. | 15:57 |
| dtantsur | there are also thoughts to integrate IPE into Ironic to simplify interactions | 15:57 |
| opendevreview | Merged openstack/ironic unmaintained/2023.1: OSSA-2025-001: Disallow unsafe image file:// paths https://review.opendev.org/c/openstack/ironic/+/949182 | 15:58 |
| dtantsur | a trivial step forward for the eventlet migration if any reviewers have time: https://review.opendev.org/c/openstack/ironic/+/949593 | 16:02 |
| cardoe | anyone else see with the latest ironicclient and OSC running openstack baremetal port show <blah> --long an error message about invalid formater? | 16:15 |
| frickler | cardoe: saw the same thing with octavia earlier, let me get the fix | 16:17 |
| frickler | https://review.opendev.org/c/openstack/python-octaviaclient/+/948981 likely ironicclient has the same issue. so much for deprecations after what was it ... 8 years? | 16:18 |
| JayF | urgh | 16:19 |
| JayF | is someone picking this up? | 16:19 |
| * JayF looking | 16:20 | |
| opendevreview | Riccardo Pittau proposed openstack/bifrost master: [WIP] use Pyhon 3.12 on centos9 https://review.opendev.org/c/openstack/bifrost/+/949449 | 16:23 |
| rpittau | I've blocked the ironicclient release https://review.opendev.org/c/openstack/releases/+/949471 in case we want to fix that before the milestone | 16:26 |
| JayF | I have a fix like 75% done | 16:34 |
| frickler | at least codesearch says that ironic should be the last victim. sorry I didn't have the idea to check that earlier, I would have expected some CI issues to come up before | 16:37 |
| opendevreview | Jay Faulkner proposed openstack/python-ironicclient master: Stop using deprecated format_* from osc_utils https://review.opendev.org/c/openstack/python-ironicclient/+/949654 | 16:38 |
| JayF | I had to modify a unit test that it doesn't seem like I should've had to modify | 16:38 |
| JayF | which is hrm-inducing | 16:38 |
| JayF | I also am going to test this manually since it'll impact visual output, seemingly | 16:38 |
| * TheJulia is reaching ENOBRAIN from board meeting | 16:39 | |
| * dtantsur is updating his document with considerations about API workers | 16:46 | |
| TheJulia | dtantsur: so, I talked with one of the operators I know who have changed threads, they went to 120-something because that matched the number of cores they had on the conductor hosts | 16:46 |
| dtantsur | This is interesting because green threads and number of CPU cores are hardly correlated | 16:48 |
| dtantsur | and I don't know if we'd want to run 120 real threads | 16:48 |
| cardoe | JayF: toss a port --node <uuid> --long in your test | 16:49 |
| JayF | cardoe: I specifically did node properties because that's one I changed :) | 16:49 |
| JayF | I'll do your port test real quick tho | 16:49 |
| JayF | I just have to create a port lol | 16:49 |
| TheJulia | dtantsur: i know... I was scratching my head some. The only thing I can really think is they are also a heavy ipmi shop | 16:51 |
| dtantsur | still, everything is running one a single core | 16:51 |
| TheJulia | yeah, but if you've got 80+ cores in wait state with ipmitool... *shrug* | 16:52 |
| TheJulia | it was what worked for them | 16:52 |
| JayF | cardoe: how did you know? | 16:52 |
| JayF | https://www.irccloud.com/pastebin/aDGlPUqY/port_extra_wtf.txt | 16:52 |
| JayF | created with ❯ openstack baremetal port create --node b5189b0e-a92e-4831-92a7-94383148985a --extra lol=cats --extra bob=son ff:ff:ff:00:11:22 | 16:53 |
| JayF | which at least in node gave me a properties | {'lol':'cats', 'bob':'son'} | 16:53 |
| TheJulia | wait, bob's the son, not the uncle?!? | 16:53 |
| * TheJulia looses a gear | 16:53 | |
| JayF | in honor or Robert B. Robertson | 16:53 |
| * TheJulia hears springs popping | 16:53 | |
| JayF | king of the bobs | 16:53 |
| TheJulia | lol | 16:53 |
| JayF | inside joke with my friends that just leaks into my fake data | 16:53 |
| TheJulia | ahh, okay | 16:57 |
| * dtantsur concludes that software was a mistake and leaves for a dinner | 17:12 | |
| dtantsur | if you wanna know why, see the "On API workers" section | 17:13 |
| opendevreview | Jay Faulkner proposed openstack/python-ironicclient master: Stop using deprecated format_* from osc_utils https://review.opendev.org/c/openstack/python-ironicclient/+/949654 | 17:19 |
| TheJulia | I guess on a plus side, the long runnings synchronous calls are few and far between | 17:21 |
| JayF | cardoe: thanks for the pointer; I compared osc-lib=3.2.0+existing code output with osc-lib=4.0.0+patch output for that call and they are identical now | 17:22 |
| JayF | cardoe: I will note that I think we're inconsistent with how we apply those formatters, so in many places with json data, we're showing {blah: blah} instead of blah=blah | 17:23 |
| JayF | I kinda want to forget about knowing ^ and just be glad I got it back to where it was lol | 17:23 |
| JayF | dtantsur: TheJulia: taking https://github.com/cherrypy/cheroot/blob/64225a3c5374a55a795c4447b80243e4897f1560/cheroot/server.py#L2165 that todo into my giant org of random oss people and trying to get them to fix :) | 17:28 |
| TheJulia | yeah, a 503 would be the most well behaved response | 17:30 |
| TheJulia | ... I wonder though how many folks front'end things today such that the resultant proxy sort of hangs | 17:30 |
| TheJulia | or just waits | 17:30 |
| JayF | then the proxy is wrong | 17:31 |
| JayF | we can be right, we can't make everyone else right :D | 17:31 |
| TheJulia | oh, often proxy handling is wrong | 17:32 |
| opendevreview | cid proposed openstack/ironic master: doc: A possible improvement around multi-arch config options https://review.opendev.org/c/openstack/ironic/+/949416 | 17:55 |
| opendevreview | Julia Kreger proposed openstack/ironic master: partial eventlet removal from some tests https://review.opendev.org/c/openstack/ironic/+/949664 | 18:15 |
| opendevreview | Julia Kreger proposed openstack/ironic master: partial eventlet removal from some tests https://review.opendev.org/c/openstack/ironic/+/949664 | 18:18 |
| TheJulia | If folks can take a glance at ^ and see if they have any issues, I can proceed with the rest of the drivers to at least lower our burden later on | 18:18 |
| JayF | I'll note the cheroot todo was added to the gr-oss upstream work backlog | 18:49 |
| JayF | I'm honestly not 100% sure what that means, but I think it means the code fairy is going to leave a pull request under our pillow at some point :D | 18:50 |
| TheJulia | excellent | 18:57 |
| JayF | if we surface any of those "no openstack context needed but python fixes needed" stuff, I can likely also get those added | 19:00 |
| JayF | Do you think that someone with minimal overall Ironic/OpenStack context could likely help with the eventlet removal directly in Ironic? | 19:00 |
| TheJulia | err, so looping call feels a bit unknown but maybe some tests taking like 9 seconds each is okay?! | 19:07 |
| TheJulia | not really sure, tbh | 19:09 |
| JayF | I mean, I'd rather have good tests that take 9 seconds each than no tests? | 19:17 |
| JayF | and plus, on the second run with stestr parallelization, won't it also do some of those in parallel? | 19:17 |
| TheJulia | oh yeah | 19:21 |
| TheJulia | I guess, ultimately, the other challenge is loopingcall use | 19:22 |
| TheJulia | like the snmp driver uses looping call to handle some aspects directly | 19:22 |
| TheJulia | it would be easier if there was a backend for it that we could directly select that was not eventlet, but it seems we're not there yet | 19:23 |
| TheJulia | :( | 19:23 |
| JayF | that's going to fix itself eventually if nobody steps up :| | 19:23 |
| TheJulia | snmp driver? | 19:23 |
| JayF | yep | 19:23 |
| TheJulia | yeah... I'll prod the one I know | 19:23 |
| * JayF double checks loopingcall implementation in oslo | 19:23 | |
| TheJulia | it looks like it is going to raise an exception on the lack of compatible backend, but... its a bit weird | 19:24 |
| JayF | https://review.opendev.org/c/openstack/oslo.service/+/945720/29/oslo_service/backend/threading/loopingcall.py | 19:24 |
| JayF | it's pretty close but still in review | 19:24 |
| TheJulia | woot | 19:24 |
| JayF | this is also what networking-baremetal eventlet migration is blocked on, fwiw | 19:24 |
| TheJulia | Some of this stuff, would be good to just change it, but I guess we can just prune the eventlet specific stuff and deal with 9 seconds | 19:25 |
| TheJulia | yeah | 19:25 |
| TheJulia | Oh, FWIW, there are a ton of wordmarks on quantum | 19:25 |
| TheJulia | like... USPTO has 2500+ | 19:25 |
| JayF | honestly I was trying to pitch a softball for someone to talk about why it changed | 19:25 |
| JayF | and gently nudge this person towards a more unique name for non-openstack reasons | 19:26 |
| TheJulia | yeah | 19:26 |
| TheJulia | I'm going to reply in a little while and denote searching marks might be a good idea, because nobody including packagers want to be stuck in trade/word mark issue land | 19:26 |
| okamitok[m] | So I sorted the issue with inspection, kolla by default doesn't have agent or inspector as `enabled_inspect_interfaces` once I got that sorted inspect works and doing a baremetal node inspect removes the ignore directive on dnsmasq so it pxe boots the ipa image and goes through that process. | 21:03 |
| okamitok[m] | Once it finishes it adds the ignore back though and when I try to do a openstack server create it kicks off the deployment, selects an instance, creates the neutron port but when the node comes up it cannot dhcp due to the ignore. Any suggestions on where to look or am I misunderstanding the flow? | 21:03 |
| JayF | "when the node comes up" meaning what exactly? | 21:09 |
| TheJulia | so there should be separate dhcp servers | 21:09 |
| JayF | does it timeout and put something in node.last_error? | 21:09 |
| TheJulia | Neutron should be launching it's own attached to then network | 21:10 |
| JayF | Or are you expecting Ironic to provide DHCP for your ACTIVE machine (we don't) | 21:10 |
| TheJulia | that dhcp server will either be in the namespace *or* supplied by OVN | 21:10 |
| JayF | https://review.opendev.org/c/openstack/ironic/+/946741 is a feature from a first-time (at least first time I've seen the name) contributor who already jumped through the hoop of moving their patch from ironic-inspector to ironic proper; if someone gets a chance we can help 'em out | 21:12 |
| okamitok[m] | <JayF> ""when the node comes up" meaning..." <- After Ironic issues the bmc power on against the bare metal node. It does eventually time out but no errors in the last_error. | 21:17 |
| okamitok[m] | When the node is powered on, my understanding would be that the ironic dnsmasq server would provide the host with a DHCP ip at which point it would then pxe and deploy the with the image provided as part of the openstack server create command? | 21:21 |
| TheJulia | okamitok[m]: That sounds like a dhcp issue. I take it your using Neutron? | 21:21 |
| TheJulia | ironic's dnsmasq dhcp server only provides addresses as it relates to inspection | 21:22 |
| okamitok[m] | Correct neutron + nova with Ironic. | 21:23 |
| TheJulia | Okay then | 21:23 |
| TheJulia | so, is Neutron using OVS or OVN ? | 21:23 |
| okamitok[m] | ovs | 21:24 |
| TheJulia | okay then | 21:25 |
| TheJulia | So your path to troubleshoot then is to find the neutron namespace which is attached to the physical network | 21:26 |
| TheJulia | I *highly* recommend looking at your neutron networks and then, on the neutron node, executing "ip netns list" and "ip netns exec <namespaceid> ip addr" to find the namespace | 21:28 |
| TheJulia | then do "ip netns exec <namespaceid> tcpdump <tcpdump args here>" to verify if your dhcp packets are making it into that namespace | 21:28 |
| TheJulia | if they are not, then that is your issue | 21:29 |
| TheJulia | if they are, then something else is going on, but tha thelps you delineate configuration versus overall state | 21:29 |
| okamitok[m] | I appreciate the guidance here, I'll go through the steps you've mentioned. The clarification on ironic dhcp only being used for inspection makes alot more sense now as I was looking through the code it appeared that was the case so I was definitely down the wrong path on troubleshooting. | 21:31 |
| TheJulia | No problem | 21:35 |
| TheJulia | Its not the easiest thing to pull apart :) | 21:35 |
| JayF | I honestly didn't even realize the ironic DHCP stuff that's built in was inspection only | 21:38 |
| TheJulia | well, if your doing anything with neutron at all, its all in neutron | 21:39 |
| okamitok[m] | I deployed ironic previously to just test the flow with bifrost, so my assumption came from there that it's also providing dhcp for deployments. | 21:39 |
| TheJulia | Ahh, bifrost configures a dhcp server to fulfill that role in the standalone case as well | 21:40 |
| opendevreview | cid proposed openstack/ironic master: Add a new configuration option, ``bootloader_by_arch`` https://review.opendev.org/c/openstack/ironic/+/949681 | 21:41 |
| JayF | TheJulia: oh, of course, we don't support mix-and-match | 21:42 |
| JayF | I get it | 21:42 |
| opendevreview | Merged openstack/sushy-tools master: Status code 501 for Unsupported/Not implemented https://review.opendev.org/c/openstack/sushy-tools/+/949500 | 21:45 |
| opendevreview | cid proposed openstack/ironic master: Add a new configuration option, ``bootloader_by_arch`` https://review.opendev.org/c/openstack/ironic/+/949681 | 21:48 |
| opendevreview | Merged openstack/ironic master: Auto-detect the RPC executor instead of hardcoding eventlet https://review.opendev.org/c/openstack/ironic/+/949593 | 21:50 |
| opendevreview | cid proposed openstack/ironic master: Add a new configuration option, ``bootloader_by_arch`` https://review.opendev.org/c/openstack/ironic/+/949681 | 21:50 |
| opendevreview | Merged openstack/ironic master: provide host_id to neutron early on https://review.opendev.org/c/openstack/ironic/+/946378 | 22:11 |
| opendevreview | cid proposed openstack/ironic master: Add a new configuration option, ``bootloader_by_arch`` https://review.opendev.org/c/openstack/ironic/+/949681 | 22:29 |
| TheJulia | JayF: I guess the question ends up being, do we drop https://github.com/openstack/ironic/blob/master/ironic/tests/unit/__init__.py#L25-L31 ... | 22:50 |
| JayF | I am of two minds. | 22:50 |
| JayF | One, it's useful to test without eventlet to know if we're making progress with removing it and not adding extra deps. | 22:51 |
| JayF | On the other hand, we've actively found bugs in some projects by ensuring eventlet is patched in beforehand | 22:51 |
| JayF | I think I'm landing on "we should stop monkey patching tests when we stop monkey patching ironic" | 22:51 |
| JayF | but that's a weakly held opinion | 22:51 |
| TheJulia | I guess I'm semi worried, and I have this running on my other monitor now... | 22:52 |
| JayF | you're going to hate my suggestion, I suspect | 22:52 |
| JayF | https://review.opendev.org/c/openstack/oslo.log/+/937729/ added a version of unit tests with a slightly different config | 22:53 |
| TheJulia | I guess I'm worried that if we leave it, we might not be able to really remove eventlet entirely until we take that hit all at once which seems awful | 22:53 |
| JayF | I'll say: I'm already shocked at how much we've been able to remove not-in-a-flag-day | 22:53 |
| JayF | I suspect we'll be really, really close, and maybe be able to change to use of greened modules explicitly, once we flip wsgi servers | 22:54 |
| TheJulia | heh, we may need to keep it | 22:54 |
| TheJulia | I think my unit test run is hung | 22:54 |
| TheJulia | so that is a sign! | 22:54 |
| JayF | yep | 22:54 |
| TheJulia | looping call is going to be a bit one | 22:54 |
| JayF | that's exactly what the early monkey_patch fixed in CI :) | 22:54 |
| JayF | that was the root cause of all of our horrible horrible migrations tests pain, remember? | 22:54 |
| JayF | and itamar finally was able to point at it with some of the eventlet improvement | 22:55 |
| TheJulia | right! | 22:55 |
| JayF | and because when we load ironic it will monkeypatch | 22:56 |
| JayF | you can't really run unit tests w/o monkeypatching, you can only run it with correct monkeypatching or wrong monkeypatching | 22:56 |
| JayF | in oslo.log I bet it was different because it supported both modes already, we don't | 22:56 |
| TheJulia | yeah | 22:56 |
| TheJulia | its going to take a while :) | 22:57 |
| JayF | I am trying to get more resources to help, the cheroot issue is a great example of something easy for me to "outsource" so to speak | 22:57 |
| JayF | if we can isolate and detail some of the other work and put a bow on it, I can try to get someone senior-pythony but maybe not openstacky to look at it | 22:57 |
| JayF | that's a hell of an "if" :) | 22:58 |
| cardoe | Random facts… MAC address is unique per Ironic for baremetal ports. Even if those devices are in two different conductor groups. | 22:58 |
| JayF | Yep, enforced by the DB. | 22:58 |
| JayF | hopefully you got a good err? | 22:58 |
| cardoe | I mean the error is fine. Off to shave another yak. | 22:59 |
| cardoe | I gotta rebase my ironic client fix on top of your fix as well. | 22:59 |
| JayF | https://opendev.org/openstack/ironic/src/branch/master/ironic/db/sqlalchemy/models.py#L256 | 23:00 |
| JayF | cardoe: alternatively get osc-lib 3.2.0 | 23:00 |
| JayF | cardoe: if it's easier to downgrade from pypi than take in a patch | 23:00 |
| cardoe | Well I’m trying to fix a bug in the client and the tests fail in gerrit without your change. | 23:11 |
| TheJulia | JayF: w/r/t the test changes, https://review.opendev.org/c/openstack/ironic/+/949664/2/ironic/tests/unit/drivers/modules/redfish/test_power.py#127 | 23:19 |
| TheJulia | you can't reliably run an assert against mocked time | 23:19 |
| TheJulia | uhh, lets see what happens this time! | 23:26 |
| opendevreview | Julia Kreger proposed openstack/ironic master: remove explicit eventlet usage around time.sleep https://review.opendev.org/c/openstack/ironic/+/949664 | 23:37 |
| TheJulia | so that seems to be behaving... | 23:37 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!